We study characteristics of receptive fields of units in deep convolutional
networks. The receptive field size is a crucial issue in many visual tasks, as
the output must respond to large enough areas in the image to capture
information about large objects. We introduce the notion of an effective
receptive field, and show that it both has a Gaussian distribution and only
occupies a fraction of the full theoretical receptive field. We analyze the
effective receptive field in several architecture designs, and the effect of
nonlinear activations, dropout, sub-sampling and skip connections on it. This
leads to suggestions for ways to address its tendency to be too small.
Описание
Understanding the Effective Receptive Field in Deep Convolutional Neural Networks
%0 Generic
%1 luo2017understanding
%A Luo, Wenjie
%A Li, Yujia
%A Urtasun, Raquel
%A Zemel, Richard
%D 2017
%K cnn
%T Understanding the Effective Receptive Field in Deep Convolutional Neural
Networks
%U http://arxiv.org/abs/1701.04128
%X We study characteristics of receptive fields of units in deep convolutional
networks. The receptive field size is a crucial issue in many visual tasks, as
the output must respond to large enough areas in the image to capture
information about large objects. We introduce the notion of an effective
receptive field, and show that it both has a Gaussian distribution and only
occupies a fraction of the full theoretical receptive field. We analyze the
effective receptive field in several architecture designs, and the effect of
nonlinear activations, dropout, sub-sampling and skip connections on it. This
leads to suggestions for ways to address its tendency to be too small.
@misc{luo2017understanding,
abstract = {We study characteristics of receptive fields of units in deep convolutional
networks. The receptive field size is a crucial issue in many visual tasks, as
the output must respond to large enough areas in the image to capture
information about large objects. We introduce the notion of an effective
receptive field, and show that it both has a Gaussian distribution and only
occupies a fraction of the full theoretical receptive field. We analyze the
effective receptive field in several architecture designs, and the effect of
nonlinear activations, dropout, sub-sampling and skip connections on it. This
leads to suggestions for ways to address its tendency to be too small.},
added-at = {2019-07-24T16:11:39.000+0200},
author = {Luo, Wenjie and Li, Yujia and Urtasun, Raquel and Zemel, Richard},
biburl = {https://www.bibsonomy.org/bibtex/273733e765b42acbbabb4e8ecbd8b3a9c/tiyuok},
description = {Understanding the Effective Receptive Field in Deep Convolutional Neural Networks},
interhash = {685400ebcb72ec49f519a8a22a41208c},
intrahash = {73733e765b42acbbabb4e8ecbd8b3a9c},
keywords = {cnn},
note = {cite arxiv:1701.04128},
timestamp = {2019-07-24T16:11:51.000+0200},
title = {Understanding the Effective Receptive Field in Deep Convolutional Neural
Networks},
url = {http://arxiv.org/abs/1701.04128},
year = 2017
}